MNRAS 437, 3103–3111 (2014) doi:10.1093/mnras/stt2041 Advance Access publication 2013 December 9 MAGIC upper limits on the GRB 090102 afterglow J. Aleksic´,1 S. Ansoldi,2 L. A. Antonelli,3‹ P. Antoranz,4 A. Babic,5 U. Barres de Almeida,6 J. A. Barrio,7 J. Becerra Gonza´lez,8 W. Bednarek,9 K. Berger,8,10 E. Bernardini,11 A. Biland,12 O. Blanch,1 R. K. Bock,6 A. Boller,12 S. Bonnefoy,7 G. Bonnoli,3 F. Borracci,6 T. Bretz,13† E. Carmona,14 A. Carosi,3‹ D. Carreto Fidalgo,7,13 P. Colin,6 E. Colombo,8 J. L. Contreras,7 J. Cortina,1 L. Cossio,2 S. Covino,3‹ P. Da Vela,4 F. Dazzi,15 A. De Angelis,2 G. De Caneva,11 B. De Lotto,2 C. Delgado Mendez,14 M. Doert,16 A. Domı´nguez,17† D. Dominis Prester,5 D. Dorner,13 M. Doro,15,18 D. Eisenacher,13 D. Elsaesser,13 E. Farina,19 D. Ferenc,5 M. V. Fonseca,7 L. Font,18 K. Frantzen,16 C. Fruck,6 R. J. Garcı´a Lo´pez,8,10 M. Garczarczyk,8 D. Garrido Terrats,18 M. Gaug,18‹ G. Giavitto,1 N. Godinovic´,5 A. Gonza´lez Munoz,1 S. R. Gozzini,11 A. Hadamek,16 D. Hadasch,20 A. Herrero,8,10 J. Hose,6 D. Hrupec,5 W. Idec,9 V. Kadenius,21 M. L. Knoetig,6 T. Kra¨henbu¨hl,12 J. Krause,6 J. Kushida,22 A. La Barbera,3 D. Lelas,5 N. Lewandowska,13 E. Lindfors,21‡ S. Lombardi,3‹ R. Lo´pez-Coto,1 M. Lo´pez,7 A. Lo´pez-Oramas,1 E. Lorenz,6,12 I. Lozano,7 M. Makariev,23 K. Mallot,11 G. Maneva,23 N. Mankuzhiyil,2 K. Mannheim,13 L. Maraschi,3 B. Marcote,24 M. Mariotti,15 M. Martı´nez,1 J. Masbou,15 D. Mazin,6 U. Menzel,6 M. Meucci,4 J. M. Miranda,4 R. Mirzoyan,6 J. Moldo´n,24 A. Moralejo,1 P. Munar-Adrover,24 D. Nakajima,6 A. Niedzwiecki,9 K. Nilsson,21‡ N. Nowak,6 R. Orito,22 A. Overkemping,16 S. Paiano,15 M. Palatiello,2 D. Paneque,6 R. Paoletti,4 J. M. Paredes,24 S. Partini,4 M. Persic,2,25 F. Prada,17,26 P. G. Prada Moroni,27 E. Prandini,15 S. Preziuso,4 I. Puljak,5 I. Reichardt,1 R. Reinthal,21 W. Rhode,16 M. Ribo´,24 J. Rico,1 J. Rodriguez Garcia,6 S. Ru¨gamer,13 A. Saggion,15 K. Saito,22 T. Saito,6 M. Salvati,3 K. Satalecka,7 V. Scalzotto,15 V. Scapin,7 C. Schultz,15 T. Schweizer,6 S. N. Shore,27 A. Sillanpa¨a¨,21 J. Sitarek,1 I. Snidaric,5 D. Sobczynska,9 F. Spanier,13 V. Stamatescu,1 A. Stamerra,4 J. Storz,13 S. Sun,6 T. Suric´,5 L. Takalo,21 F. Tavecchio,3 P. Temnikov,23 T. Terzic´,5 D. Tescaro,8,10 M. Teshima,6 J. Thaele,16 O. Tibolla,13 D. F. Torres,20,28 T. Toyama,6 A. Treves,19 M. Uellenbeck,16 P. Vogler,12 R. M. Wagner,6 Q. Weitzel,12 F. Zandanel,17§ R. Zanin,24 A. Bouvier,29 M. Hayashida,30 H. Tajima30,31 and F. Longo32 1 IFAE, Edifici Cn., Campus UAB, E-08193 Bellaterra, Spain 2Universita` di Udine, and INFN Trieste, I-33100 Udine, Italy 3INAF National Institute for Astrophysics, I-00136 Rome, Italy 4Universita` di Siena, and INFN Pisa, I-53100 Siena, Italy  E-mail: angelo.antonelli@oa-roma.inaf.it (LAA); alessandro.carosi@oa-roma.inaf.it (AC); stefano.covino@oa-roma.inaf.it (SC); markus.gaug@uab.cat (MG); saverio.lombardi@oa-roma.inaf.it (SL) † Present address: Ecole polytechnique fe´de´rale de Lausanne (EPFL), Lausanne, Switzerland. ‡ Present address: Finnish Centre for Astronomy with ESO (FINCA), University of Turku, Finland. § Present address: GRAPPA Institute, University of Amsterdam, NL-1098 XH Amsterdam, the Netherlands. C© 2013 The Authors Published by Oxford University Press on behalf of the Royal Astronomical Society at U niversity of Turku on February 2, 2015 http://m nras.oxfordjournals.org/ D ow nloaded from 3104 J. Aleksic´ et al. 5Croatian MAGIC Consortium, Rudjer Boskovic Institute, University of Rijeka and University of Split, HR-10000 Zagreb, Croatia 6Max-Planck-Institut fu¨r Physik, D-80805 Mu¨nchen, Germany 7Universidad Complutense, E-28040 Madrid, Spain 8Inst. de Astrofı´sica de Canarias, E-38200 La Laguna, Tenerife, Spain 9University of Ło´dz´, PL-90236 Lodz, Poland 10Depto. de Astrofı´sica, Universidad de La Laguna, E-38206 La Laguna, Spain 11Deutsches Elektronen-Synchrotron (DESY), D-15738 Zeuthen, Germany 12ETH Zurich, CH-8093 Zurich, Switzerland 13Universita¨t Wu¨rzburg, D-97074 Wu¨rzburg, Germany 14Centro de Investigaciones Energe´ticas, Medioambientales y Tecnolo´gicas, E-28040 Madrid, Spain 15Universita` di Padova and INFN, I-35131 Padova, Italy 16Technische Universita¨t Dortmund, D-44221 Dortmund, Germany 17Inst. de Astrofı´sica de Andalucı´a (CSIC), E-18080 Granada, Spain 18Unitat de Fı´sica de les Radiacions, Departament de Fı´sica, and CERES-IEEC, Universitat Auto`noma de Barcelona, E-08193 Bellaterra, Spain 19Universita` dell’Insubria, Como, I-22100 Como, Italy 20Institut de Cie`ncies de l’Espai (IEEC-CSIC), E-08193 Bellaterra, Spain 21Finnish MAGIC Consortium, Tuorla Observatory, University of Turku and Department of Physics, University of Oulu, Finland 22Japanese MAGIC Consortium, Division of Physics and Astronomy, Kyoto University, Japan 23Inst. for Nucl. Research and Nucl. Energy, BG-1784 Sofia, Bulgaria 24Universitat de Barcelona (ICC/IEEC), E-08028 Barcelona, Spain 25INAF/Osservatorio Astronomico and INFN, I-34143 Trieste, Italy 26Instituto de Fisica Teorica, UAM/CSIC, E-28049 Madrid, Spain 27Universita` di Pisa, and INFN Pisa, I-56126 Pisa, Italy 28ICREA, E-08010 Barcelona, Spain now at Department of Physics & Astronomy, UC Riverside, CA 92521, USA 29Santa Cruz Institute for Particle Physics, University of California, Santa Cruz, CA 95064, USA 30W. W. Hansen Experimental Physics Laboratory, Kavli Institute for Particle Astrophysics and Cosmology, Department of Physics and SLAC National Accelerator Laboratory, Stanford University, Stanford, CA 94305, USA 31Solar-Terrestrial Environment Laboratory, Nagoya University, Nagoya 464-8601, Japan 32Dipartimento di Fisica, Universita‘ di Trieste and INFN Trieste, I-34127 Trieste, Italy Accepted 2013 October 20. Received 2013 October 9; in original form 2013 August 27 ABSTRACT Indications of a GeV component in the emission from gamma-ray bursts (GRBs) are known since the Energetic Gamma-Ray Experiment Telescope observations during the 1990s and they have been confirmed by the data of the Fermi satellite. These results have, however, shown that our understanding of GRB physics is still unsatisfactory. The new generation of Cherenkov observatories and in particular the MAGIC telescope, allow for the first time the possibility to extend the measurement of GRBs from several tens up to hundreds of GeV energy range. Both leptonic and hadronic processes have been suggested to explain the possible GeV/TeV counterpart of GRBs. Observations with ground-based telescopes of very high energy (VHE) photons (E > 30 GeV) from these sources are going to play a key role in discriminating among the different proposed emission mechanisms, which are barely distinguishable at lower energies. MAGIC telescope observations of the GRB 090102 (z = 1.547) field and Fermi Large Area Telescope data in the same time interval are analysed to derive upper limits of the GeV/TeV emission. We compare these results to the expected emissions evaluated for different processes in the framework of a relativistic blastwave model for the afterglow. Simultaneous upper limits with Fermi and a Cherenkov telescope have been derived for this GRB observation. The results we obtained are compatible with the expected emission although the difficulties in predicting the HE and VHE emission for the afterglow of this event makes it difficult to draw firmer conclusions. Nonetheless, MAGIC sensitivity in the energy range of overlap with space-based instruments (above about 40 GeV) is about one order of magnitude better with respect to Fermi. This makes evident the constraining power of ground-based observations and shows that the MAGIC telescope has reached the required performance to make possible GRB multiwavelength studies in the VHE range. Key words: radiation mechanisms: non-thermal – gamma-ray burst: general. at U niversity of Turku on February 2, 2015 http://m nras.oxfordjournals.org/ D ow nloaded from MAGIC observation of GRB 090102 afterglow 3105 1 IN T RO D U C T I O N Since the discovery of gamma-ray bursts (GRBs) in the late 1960s (Klebesadel, Strong & Olson 1973), these energetic and mysterious phenomena have been targets of large observational efforts. The discovery of their afterglow in late 1990s (Costa et al. 1997; Van Paradijs et al. 1997) provided a great boost in GRB studies at all wavelengths. The wealth of available information put severe con- straints on the various families of interpretative scenarios, showing an unexpected richness and complexity of possible behaviours (see e.g. Gehrels, Ramirez-Ruiz & Fox 2009). The first observations at MeV–GeV energies with the Energetic Gamma-Ray Experiment Telescope (EGRET) on board the Compton Gamma-Ray Obser- vatory (Hurley et al. 1994; Dingus 1995), showed that the high energy (HE: 1 MeV–30 GeV) and very high energy range (VHE: 30 GeV–30 TeV) can be powerful diagnostic tools for the emission processes and physical conditions of GRBs. The launch of Fermi (Band et al. 2009), with its Large Area Telescope (LAT; Atwood et al. 2009b), showed that, at least for the brightest events, GeV emission from GRBs is a relatively common phenomenon (Granot et al. 2010). However, a satisfactory interpretative framework of the GeV emission is still lacking. In this context, ground-based imag- ing atmospheric Cherenkov telescopes (IACTs), such as MAGIC,1 H.E.S.S.2 and VERITAS,3 despite the reduced duty cycle of ground- based facilities, provide access to the ∼100 GeV to TeV energy in- terval for GRB observations. Furthermore, the energy range down to ∼80 GeV, which was accessible almost exclusively with space- based instruments, has been opened to ground-based observations by the MAGIC observatory (Aliu et al. 2008; Schweizer et al. 2010). Together with the multiwavelength coverage provided by the LAT instrument, this makes possible the complete coverage of the 1– 100 GeV energy range with the advantage, in the VHE domain, of an increase of ∼2–3 order of magnitude in the sensitivity relative to space-based instruments. Moreover, the low-energy trigger thresh- old of MAGIC makes less relevant the effect of the source distance. The flux above ∼100 GeV is attenuated by pair production with the lower energetic (optical/IR) photons of the diffuse Extragalactic Background Light (EBL; Nikishov 1962; Gould & Schreder 1966). The resulting cosmic opacity to VHE gamma-rays heavily affects Cherenkov observations, especially for GRBs which are sources with an average redshift slightly larger than 2 (Fynbo et al. 2009). Therefore, the higher the redshift, the lower the likelihood of detec- tion at a given energy [i.e. about a 90 per cent of flux reduction at 100 GeV for a z = 2 source following Domı´nguez et al. (2011a)]. In addition, the transient and unpredictable nature of GRBs makes it difficult for large ground-based instruments such as IACTs to point them rapidly enough to catch the prompt emission and the early afterglow phases, when these sources are expected to be observable at high energies (see e.g. Covino et al. 2009a, for a discussion about IACTs perspectives for GRB observations). MAGIC has the ad- vantage, compared to the other IACTs, in its low-energy sensitivity and pointing speed (e.g. Garczarczyk et al. 2009). Several attempts to observe GRB emission have been discussed (Albert et al. 2006, 2007; Aleksicˇ et al. 2010). In all cases, only upper limits (ULs) have been derived. Similar results have also been reported by other IACTs (Tam et al. 2006; Aharonian et al. 2009; Acciari et al. 2011). As 1 http://wwwmagic.mppmu.mpg.de/index.en.html 2 http://www.mpi-hd.mpg.de/hfm/HESS 3 http://veritas.sao.arizona.edu discussed above, the two most limiting factors are the high redshift of the source and the delay of the observation. In this paper, we report and discuss the MAGIC observation of GRB 090102, a GRB at a redshift about 1.5 observed at low zenith angle and good weather conditions. These observations permitted data-taking with an energy threshold of about 30 GeV. However, no gamma-ray signal was detected, and hence only ULs could be derived. Section 2 gives general information about GRB 090102. In Sec- tions 3 and 4, we discuss the MAGIC and LAT data sample, re- spectively, while Sections 5 and 6 introduce and develop the inter- pretative scenario. In Section 7, we evaluate the effect of the EBL absorption on the lowest energy bins allowed by our observation and finally, we discuss our results in a general theoretical scenario in the last section. We assume a  cold dark matter cosmology with m = 0.27,  = 0.73 and h0 = 0.71. At the redshift of the GRB, the proper distance is ∼4.5 Gpc (∼1.38 × 1028 cm). Throughout this paper the convention Qx = Q/10x has been adopted in CGS units. 2 G R B 0 9 0 1 0 2 GRB 090102 was detected and located by the Swift satellite (Gehrels et al. 2004) on 2009 January 2 at 02:55:45 UT (Mangano et al. 2009b) and also by the Fermi Gamma-ray Burst Monitor (GBM) detector. The prompt light curve was structured in four partially overlap- ping peaks (Sakamoto et al. 2009) for a total T90 of 27.0 ± 2.0 s. Since the burst was also detected by Konus Wind (Golenetskii et al. 2009) and Integral (Mangano et al. 2009a), it has been possible to obtain a very good reconstruction of the prompt emission spec- tral parameters. The time-averaged spectrum can be modelled with the classical Band function (Band et al. 1993) with peak energy Epeak = 451 +73−58 keV and a total fluence in the 20 keV–2 MeV range of 3.09 +0.29−0.25 × 105 erg cm−2 (Golenetskii et al. 2009). Early op- tical follow-up measurements were performed by many groups like TAROT (Klotz et al. 2009) at T0+40.8 s, the REM robotic telescope at T0+53 s (Covino, D’Avanzo & Antonelli 2009b) and GROND telescope (Afonso et al. 2009) at T0+2.5 h. Optical spec- troscopy was rapidly obtained with the NOT telescope by De Ugarte Postigo et al. (2009a). They found evidence of several absorp- tion metal lines, including Fe II, Mg II, Mg I, Al II, Al III and C IV, at a common redshift of z = 1.547. The resulting isotropic en- ergy value Eiso = 5.75 × 1053 erg and the rest-frame peak energy Epeak = 1149 +186−148 keV are in good agreement with the Amati rela- tion (Amati, Frontera & Guidorzi 2009). The multiwavelength light curve is shown in Fig. 1 in which, data in the R and H band corre- spond to a rest-frame UV and optical emission, respectively. X-ray data are the unbinned Swift X-Ray Telescope (XRT) and Burst Alert Telescope (BAT) data in the 0.5–10 keV. According to Gendre et al. (2010), it is very difficult to model the whole afterglow in a standard scenario (see the next section). Moreover, it showed a distinct be- haviour in the optical and in X-rays. The X-ray light curve showed an uninterrupted decay from about 400 s from the GRB onset up to 5 × 106 s, when Swift ceased observations of the event. The optical light curve, monitored from several tens of seconds to slightly more than a day from the T0, showed a steep-to-shallow behaviour with a break at about 1 ks. Before the break, the optical flux decay index is α1 = 1.50 ± 0.06 while the index becomes α2 = 0.97 ± 0.03 after the break, steeper and flatter, respectively, when compared to the simultaneous X-ray emission. This behaviour strongly resembles that showed by GRB 061126 (Gomboc et al. 2008; Perley et al. 2008) and GRB 060908 (Covino et al. 2010). at U niversity of Turku on February 2, 2015 http://m nras.oxfordjournals.org/ D ow nloaded from 3106 J. Aleksic´ et al. Figure 1. Light curve for GRB 090102. Data in the R band (red points) were taken from the TAROT, REM, NOT, GROND and Palomar telescopes. The early near-infrared H band (blue points) are from the observations of the REM telescope. All magnitudes are expressed in the Vega system. X-ray data are the unbinned Swift/XRT and BAT data in the 0.5–10 keV (green and magenta point, respectively). The MAGIC observation window is also plotted. R and H data from Gendre et al. (2010). XRT and BAT data retrieved from http://www.swift.ac.uk/burst_analyser (Evans et al. 2010). Other observations were performed at later time by GROND (T0+2.5 h; Afonso et al. 2009), Palomar (T0+50 min; Cenko, Rau & Salvato 2009) and IAC80 (T0+19.2 h; De Ugarte Postigo, Blanco & Castro-Tirando 2009b) telescopes while during the following days, the NOT (Malesani et al. 2009) and HST (Levan et al. 2009) provided the detection of the host galaxy. In the radio energy band, the VLA (Chandra & Frail 2009) and the Westerbork Synthesis Ra- dio Telescope (Van der Horst, Wijers & Kamble 2009) performed follow-up observation at 8.46 and 4.9 GHz with no afterglow de- tection and UL evaluation. A detailed discussion of the follow-up observations for this burst can be found in Gendre et al. (2010). 3 M AG IC FO L L OW-U P O B SE RVAT IO N A N D A NA LY S I S The MAGIC telescope located at Roque de los Muchachos (28.◦75 N, 17.◦89 W, La Palma, Canary Islands) performed a follow-up mea- surement of GRB 090102. The data presented in this paper were taken when MAGIC was operating as a single telescope. The MAGIC telescope was autonomously repointed and started the ob- servations at T0+255 s, following the GRB alert from Fermi-GBM. Later on, the shift crew operating the telescope realized that the GBM coordinates (RA: 08h35m06s; Dec.: 37◦16′48′ ′) differed from the BAT coordinates (RA: 08h33m02s; Dec.: 33◦05′29′ ′) by more than 4◦. Consequently, the telescope was repointed to the BAT coor- dinates and re-started observations by T0+1161 s. After this burst, the alert system was modified to cope with this situation. First data runs were taken at very low zenith angles from 5◦ reaching 52◦ at the end of data taking at 06:54:01 UT after 13 149 s of observation. MAGIC ULs above 80 GeV have already been published for this GRB (Gaug et al. 2009a), while results and scientific discussion about a subsequent dedicated analysis focused in the low-energy band (Gaug et al. 2009b) will be presented here. To ensure the low- est energy threshold, only data taken with zenith distance <25◦, corresponding to the first 5919 s of observation (data subsample up to 04:53:32 UT) have been taken into account during this analysis. By employing the MAGIC-1 sum trigger system (Aliu et al. 2008), an Table 1. MAGIC-I 95 per cent confidence level ULs for the afterglow emission of GRB 090102. The val- ues correspond to the first 5919 s of observation from 03:14:52 to 04:53:32 UT. α Bins central energy was evaluated applying all analysis cuts to MC simulations. β Statistical significance of the excess events observed by MAGIC. E bin 〈E〉α σ β Average flux limits (GeV) (GeV) (erg cm−2 s−1) 25–50 43.9 0.83 8.7 × 10−10 50–80 57.3 − 0.30 1.5 × 10−10 80–125 90.2 1.09 3.1 × 10−10 125–175 137.2 0.51 2.2 × 10−10 175–300 209.4 0.90 1.6 × 10−10 300–1000 437.6 − 0.48 0.3 × 10−10 analysis threshold of around 30 GeV is achieved, which is evaluated from Monte Carlo (MC) simulations. In order to accurately estimate the background from hadronic atmospheric showers, an OFF data sample was taken one night later with the telescope pointing close to the burst location and in the same observational conditions and instrument setup. Data were analysed using the MAGIC Analysis and Reconstruction Software (MARS; Albert et al. 2008; Aliu et al. 2009a) and processed using the standard Hillas parameters (Hillas 1985). Gamma/hadron separation and energy estimation were per- formed using a multidimensional classification method (Random Forest; Breiman 2001) while arrival directions of the gamma pho- tons is reconstructed using the DISP algorithm (Fomin et al. 1994). The alpha parameter is then used to evaluate the significance of the signal in six energy bins. In spite of the low-energy analysis threshold, no significant excess of gamma-ray photons have been detected from a position consistent with GRB 090102. Differential ULs assuming a power-law gamma-ray spectrum with spectral in- dex of  = −2.5 and using the method of Rolke, Lo´pez & Conrad (2005) were evaluated with a 95 per cent confidence level (CL) and 30 per cent estimation of systematic uncertainties and are reported in Table 1 and Fig. 2. 4 L AT O B S E RVAT I O N A N D A NA LY S I S The Fermi observatory is operating in a sky survey mode and the Swift localization of GRB 090102 was observable by the LAT in- strument approximately 3300 s after trigger and remained within the LAT field of view ( boresight  60◦) for a duration of ∼2300 s. We analysed the Fermi-LAT data using the Science Tools 09- 30-01 with Pass7V6 ‘Source’ event class. We used the publicly available models for the Galactic and isotropic diffuse emissions, gal 2yearp7v6 trim v0.f its and iso p7v6source.txt , that can be retrieved from the Fermi Science Support Center.4 No signifi- cant excess was found in this observation, so we computed ULs in three different energy bands: [0.1–1 GeV], [1–10 GeV] and [10– 100 GeV]. We first fit the broad energy range (from 0.1 to 100 GeV) using the unbinned likelihood analysis, which was then used to constrain the background model. Then we froze the normalizations of the isotropic and Galactic diffuse templates, and independently fit the source in the three different energy bands, using the unbinned profile likelihood method to derive 95 per cent LAT ULs. The following UL values were derived for the 4 http://fermi.gsfc.nasa.gov/ssc/data/access/lat/BackgroundModels.html at U niversity of Turku on February 2, 2015 http://m nras.oxfordjournals.org/ D ow nloaded from MAGIC observation of GRB 090102 afterglow 3107 Figure 2. SSC modelled emission during the afterglow of GRB 090102. Blue triangles are 95 per cent CL ULs derived by MAGIC for low-energy (LE) analysis. The relatively more constraining UL in the 50–80 GeV is due to a negative significance energy bin. For comparison, the regular energy range MAGIC ULs (Gaug et al. 2009a) are also reported in light grey. The red triangles report the Fermi-LAT 95 per cent CL ULs. The purple and black curves depict the expected energy flux according to the GRB afterglow model described in Sections 6 and 5. Physical parameters are e = 0.1, B = 0.01, E52 = 4.5 and T = T0 + 4 ks at a redshift z = 1.547. The shaded region shows the uncertainty in the EBL absorption, as prescribed in Domı´nguez et al. (2011a). [0.1–1 GeV], [1–10 GeV], [10–100 GeV] energy ranges, respec- tively: 2.73 × 10−10, 4.58 × 10−10, 3.45 × 10−9 erg cm−2 s−1 and are depicted in Fig. 2. These ULs are more constraining than the ones reported in Inoue et al. (2013). The reason for that is the usage of P7V6 ‘Source’ instead of P6V3 ‘Diffuse’, and also the usage of a different procedure to parametrize the diffuse background in the three differential energy bins. Even if observed with a consider- able time delay, the achieved energy threshold of MAGIC permits a better overlap with LAT in the GeV range when compared with previous results on GRB by MAGIC and other IACTs. Thus, it has been possible to derive simultaneous ULs with a complete cov- erage of the energy range from 0.1 GeV up to TeV using MAGIC and Fermi-LAT. Furthermore, it is worth stressing that, in the energy range where the two instruments overlap (range [25–100 GeV]), the ULs derived by MAGIC are about one order of magnitude lower than those from Fermi-LAT. 5 T H E L OW-E N E R G Y S C E NA R I O In a commonly accepted scenario (see e.g. Me´sz´aros 2006, for a review), GRB dynamics during the prompt phase are governed by relativistic collisions between shells of plasma emitted by a central engine (internal shocks). Similarly, the emission during the after- glow is thought to be connected to the shocks between these ejecta with the external medium (external shocks). Several non-thermal mechanisms, indeed, have been suggested as possible sources of HE and VHE5 photons. They include both leptonic and hadronic processes (see e.g. for a review Zhang & Me´sz´aros 2001; Gupta & Zhang 2007; Fan & Piran 2008; Ghisellini 2010). In the most 5 GRBs show their phenomenology mainly in the X-ray and soft γ -ray energy band (1 keV–1 MeV). To avoid confusion with the Fermi-LAT and IACT operational energy range (>20 MeV and >25 GeV, respectively), we will refer to the former as a ‘low-energy’ range. plausible scenario, electron synchrotron radiation is the dominant process in the low-energy regime. Within this scenario, the GRBs spectra are usually approximated by a broken power law in which the relevant break energies are the minimum injection νm and the cooling νc. The first one refers to emission frequency of the bulk of the electron population (where most of the synchrotron emission occurs), while the cooling frequency identifies where electrons ef- fectively cool. Both are strongly dependent on the microphysical parameters used to describe the GRB environment and, for a con- stant density n of the circumburst diffuse interstellar medium, they are given by (Zhang & Me´sz´aros 2001) νm = 8.6 × 1017 ( p − 2 p − 1 )2 ( e ζe )2 t −3/2 h E 1/2 52 1/2 B (1 + z)1/2 [Hz] (1) νc = 3.1 × 1013 (1 + Ye)−2 −3/2B E−1/252 n−1t−1/2h (1 + z)1/2 [Hz], (2) where e and B are the energy equipartition parameter for electrons and magnetic field, E52 is the energy per unit solid angle, th is the observer’s time in hours, ζ e is fraction of the electrons that enter in the acceleration loop and Ye is the ratio between synchrotron and Inverse Compton (IC) cooling time, known as Compton factor (see e.g. Panaitescu & Kumar 2000; Sari & Esin 2001). As a matter of fact, we have explicitly assumed that the contribution of the Comp- ton scattering is not negligible in the afterglow at the considered time and, as a consequence, the cooling break is reduced by a factor (1+Ye). It is important to remark that the change in slope of the optical decay observed in GRB 090102 suggests that the standard model cannot adequately describe the dynamics of this event. The steep-to-shallow behaviour could be interpreted as due to a termina- tion shock, locating the end of the free-wind bubble generated by a massive progenitor at the position of the optical break. However, it is also possible to hypothesize that the early steeper decay is simply due to the superposition of the regular afterglow and a reverse shock present only at early times. It is not our purpose to analyse and dis- cuss the several physical scenarios that are proposed to describe the afterglow, so we continue to model the burst emission assuming the afterglow could be described in the standard context of a relativistic shock model. 6 M O D E L I N G T H E V H E E M I S S I O N Any attempt to a meaningful modelling of the possible VHE emis- sion component, both during the prompt emission and the after- glow, must rely on information coming from the low energies (see e.g. Aleksicˇ et al. 2010). At the same time, the modelling of the low-energy afterglow can furthermore help in limiting the intrinsic degeneracy or even, to some extent, arbitrariness in the choice of the various possible HE and VHE afterglow parameters. Following Gendre et al. (2010), we assume that the cooling frequency at the time of MAGIC observation is located between optical and X-ray bands. Thus, we can estimate the slope of the energy particles dis- tributions which is correlated with the optical decay index. With the observed optical spectral index of 0.97 ± 0.03 (Gendre et al. 2010), we obtain a value for p from the relation 43 (p − 1) = 0.97 of p = 2.29 ± 0.04 in good agreement with numerical simula- tions which suggest a value of p ranging between 2.2 and 2.3 (Achterberg et al. 2001; Vietri 2003). We will assume that at the time of the MAGIC observation, the outflow expands into a dif- fuse medium with a constant density of the order of n ∼ 1 cm−3, at U niversity of Turku on February 2, 2015 http://m nras.oxfordjournals.org/ D ow nloaded from 3108 J. Aleksic´ et al. Table 2. Spectrum break energies for the different considered pro- cesses. The value of the expected SSC emission in the first MAGIC energy bin (∼40 GeV) is also shown with and without considering the EBL absorption. We refer to Zhang & Me´sz´aros (2001) for the numerical results presented in this paper. Synchrotron (e) SSC Synchrotron (p) Em  0.6 eV Esscm  1.1 MeV Epm  10−8eV Ec  4.1 eV Esscc  47 MeV Epc  140 TeV Emax  207 MeV EKN  60 TeV Epmax  1.7 MeV ≈5 × 10−11 ≈1.1 × 10−10 ≈4 × 10−17 – 4.3 × 10−11 (3.4 × 10−11) – and we will further assume that all electrons are accelerated in the shocks (ζ e ∼ 1). At the same time, from the available data, we can only constrain the values of B and e. Assuming that the optical light-curve time break (Melandri et al. 2010) is less than the start time of the shallow decay phase (Tbreak  103 s), we obtain 0.04  e  0.2 and 7 × 10−4  B  0.05 which only barely fix the B, e values. We thus assume, within these limits, B ∼ 0.01 and e ∼ 0.1 which correspond to typical values for the late afterglow (Panaitescu & Kumar 2002; Yost et al. 2003). The most plausible process producing VHE photons is the IC mechanism in the variant of Synchrotron Self Compton (SSC; Sari & Esin 2001; Zhang & Me´sz´aros 2001). Within this process, the low-energy photons pro- duced in the standard synchrotron emission are the seed photons that are pushed into the VHE band by IC scattering. Similarly to the previous case, the predicted SSC spectrum is characterized by the two typical frequencies νsscm = γ 2mνm and νsscc = γ 2c νc, where γ m and γ c are the Lorentz factor for the electrons of frequencies νm and νc. Since electrons are ultrarelativistic (γm,c ∼ 103), SSC radiation can easily reach the GeV–TeV domain. Following Zhang & Me´sz´aros (2001) and according to equations (1) and (2), we have νsscm = 1.3 × 1022 ( e ζe )4 ( p − 2 p − 1 )4 × E3/452 n−1/4t−9/4h 1/2B (1 + z)5/4 [Hz] (3) νsscc = 1.2 × 1025(1 + Ye)−4 × E−5/452 n−9/4 −7/2B t−1/4h (1 + z)−3/4 [Hz], (4) while the expected maximum flux density is (Zhang & Me´sz´aros 2001) F sscν,max = 17 ζ 2e (nE52)5/4t1/4h (1 + z)3/4 [nJy]. (5) Basically, the new spectral feature has the same shape of the underlying synchrotron component with a new break in the spectrum (EKN) due to the decreasing of the IC cross-section with energy (Fragile et al. 2004). However, this cut-off is found to be above few tens of TeV in our case, securing that MAGIC ULs stay below this limit. The relevant break energies for the assumed model are summarized in Table 2. We also consider proton synchrotron emission as a hadronic- originated component (Bo¨ttcher & Dermer 1998; Pe’er & Waxman 2005). Although protons are poor emitters with respect to the elec- trons due to their high mass, they can be accelerated in internal or external shocks in the same way as electrons producing synchrotron radiation. However, the peak flux for the two particle emission com- ponents is determined by the mass ratio ≈ me mp which implies that hadronic component is usually much smaller with respect to the electron emission. Nevertheless, while electrons cool quickly, pro- tons cooling times are much longer since νc,p νc,e = ( 1 + Ye 1 + Yp )2 ( mp me )6 , (6) where Yp = σpγσp,T Ye and σpγ , σp,T are the proton–gamma interaction and Thomson cross-section respectively. We note that since σpγ σp,T , Yp 1 and proton’s energy is mostly lost in proton–gamma interaction rather than synchrotron emission. However, equation (6) implies that in the same cases proton synchrotron emission can exceed the electron component in the MAGIC energy band. In Section 8, we will briefly discuss the relative importance of the above described emission components. 7 E B L AT T E N UAT I O N Gamma-ray absorption by pair production with EBL plays a key role in VHE astronomy since it significantly limits the IACTs capa- bility in detecting sources at redshift z > 1. The optical depth τ is strictly connected to the light content of the Universe and the source distance. In the past years, several EBL models have been proposed providing a wide range of values for τ from 1 up to 6 for a z ∼ 1 source at 100 GeV (see e.g. Kneiske et al. 2004; Stecker, Malkan & Scully 2006; Stecker & Scully 2008), which gives an attenuation in the expected flux ranging between 1/3 and to more than 1/100. However, the more recent EBL models (Franceschini, Rodighiero & Vaccari 2008; Domı´nguez et al. 2011a), although based on different assumptions, are converging to stable results. Within this context, Domı´nguez et al. (2011a) have used real data on the evolution of the galaxy population taken from the All-wavelength Extended Groth Strip International Survey (AEGIS) catalogue to evaluate EBL in- tensity for a wide range of redshift. The reliability of the results have been tested on the three most distant object observed by MAGIC (Domı´nguez et al. 2011b). Moreover, the EBL intensity evaluated using this model matches the minimum level allowed by galaxy counts which leads to the highest transparency of the universe to VHE gamma-rays. We used the model of Domı´nguez et al. (2011a) to evaluate the EBL absorption obtaining a value for τ of 0.218+0.075−0.041 at about 40 GeV. This gives an attenuation of the flux at the same energy of the order of ∼20 per cent, a value that does not signif- icantly compromise detection capability of MAGIC. However, the optical depth increases quickly with energy reaching the values of ∼1.5 and ∼14.4 at 100 and 500 GeV, respectively, and this makes necessary to lower the energy threshold of the observation. In the case of GRB 090102, MAGIC shows its capability to perform ob- servation at very low energy limiting the gamma absorption even for moderate redshift sources. 8 D I SCUSSI ON Although only ULs have been obtained, the possibility of having simultaneous observations with Fermi-LAT in the energy range 0.1–100 GeV and MAGIC in the energy range that starts at 25 GeV (hence overlapping with LAT) make the GRB 090102 a good case study in spite of its relatively high redshift. However, it has to be remarked that GRB 090102 can be considered as a common GRB in terms of both energetics and redshifts. Higher expected fluxes can be foreseen in the case of more energetic events that are not so rare accordingly to Fermi results. We have used the equations of relativistic shock model in order to predict, in a reliable way, at U niversity of Turku on February 2, 2015 http://m nras.oxfordjournals.org/ D ow nloaded from MAGIC observation of GRB 090102 afterglow 3109 Figure 3. Modeled SED in hadronic-dominated scenario for GRB af- terglow. Used parameters are E52 = 103, Tobs = T0 + 4 ks, e = 10−3, B = 0.01, n = 100 cm−3. the expected VHE emission in the LAT and MAGIC energy range. From numerical results, it is evident that for the chosen parameters and at our observation time, leptonic components are the dominant mechanisms from the low to the VHE. Following equation (6), the cooling frequency for protons is usually located well above the VHE range (>TeV) and this make the process potentially interest- ing for MAGIC observations. However, to make the two emissions comparable in the low-energy regime, and the proton component to dominate the leptonic one at high energies, a fine tuning in the parameters choice is needed implying e p ≈ me mp ≈ 10−3. Similar re- sults can be obtained for the IC component. In both cases, however, a higher total energy release of ≈1055 erg and a circumburst density medium of ≈100 cm−3 are needed to maintain the low-energy flux at the observed level. This makes the possibility of observing the hadronic emission component with the MAGIC telescope unrealis- tic, at least for a canonical model. A sketch of the scenario described above is shown in Fig. 3. Here, we did not take into account other hadronic-induced pro- cesses such as π0 decay (Bo¨ttcher & Dermer 1998). However, it has been shown that they could have a non-negligible effect at higher en- ergies. For our parameters, the SSC process looks the most reliable mechanism in the VHE range. Indeed, we obtain Esscm ∼ 1 MeV and Esscc ∼ 50 MeV. We conclude that MAGIC observation (>40 GeV) was carried out in the spectral energy distribution (SED) region where νFν ∝ ν(2 − p)/2 (Wei & Fan 2007) so that it is possible to evaluate the expected SSC emission. Following Zhang & Me´sz´aros (2001), νFν = νFν,maxνc,ssc1/2νm,ssc(p−1)/2ν(2−p)/2, (7) which gives (for MAGIC first energy bin and taking into account the EBL absorption) νFν,ssc(40 GeV) ≈ 3.4 × 10−11 erg cm−2 s−1. This result lies about one order of magnitude below the corresponding ULs. However, a change in the microphysical parameters can influ- ence the VHE emission giving scenarios with substantially higher flux. One of the most critical variables is the intensity of the mag- netic field, which influences the relative importance between syn- chrotron and SSC emission. It is of particular importance for this event, since the most striking observational feature of GRB 090102 was the observation of ∼10 per cent polarization in the optical at early times (about 3 min after the GRB; Steele et al. 2009). One of the most plausible interpretations is that the outflow generating the GRB is driven by a large-scale ordered magnetic field, which gener- Figure 4. The expected SSC emission at different time. T0 + 0.8 ks (purple curve), T0 + 2 ks (yellow curve), T0 + 10 ks (orange curve). In the latter case, the corresponding VERITAS 99 per cent ULs on ∼5000 s of observation have been plotted (Acciari et al. 2011) for comparison. ates polarized optical synchrotron emission in the optical observable during the reverse shock phase (Steele et al. 2009). Large values of the magnetic field affect in a significant way the HE emission since it reduces the importance of the IC component. However, the regular forward-shock emission should not be affected by this ordered mag- netic field (Covino 2007; Mundell et al. 2007). The time-scale of the MAGIC observations are indeed likely late enough not to require this further parameter in the modelling. Such a delay, in associa- tion with the moderate source distance, militate against performing constraining observations with MAGIC. Indeed, we estimated that νFν ∝ t−1.2. This implies that lowering the temporal delay of the observations can make the expected emission higher by one order of magnitude as illustrated in Fig. 4 where the expected SSC emission at different time is showed. 9 FUTURE PROSPECTS Catching VHE signal from GRBs is one of the primary target of the MAGIC telescope and future IACTs like the Cherenkov Telescope Array (CTA). Our estimates show that for this particular GRB, MAGIC follow-up observations made within the first 1–2 min from the trigger time would have the potential to detect the VHE com- ponent or at least to evaluate constraining ULs (see Fig. 5). This demonstrates both the capabilities of the system and the necessity of a fast-response observations. As GeV emission is found to be relatively common in Fermi GRBs (see e.g Abdo et al. 2009c), the unique opportunity of having simultaneous follow-up with LAT and the MAGIC telescope will make accessible the end of the electromagnetic spectrum of GRBs and will have an important role in constraining different emission mechanisms and the space parameters. Moreover, the recent tech- nical improvement of the MAGIC stereo system (Aleksicˇ et al. 2012) will bring an improvement in the instrument sensitivity in its low-energy range. The steeper decay of the flux makes in any case difficult late time (>200 s) detections for such moderate high- redshift event (see Fig. 5). On the other hand, such a time-scale is well within the pointing capabilities of the present generation of IACTs (e.g. MAGIC) that are able to perform follow-up measure- ments within few hundreds of seconds. Basing on the preliminary at U niversity of Turku on February 2, 2015 http://m nras.oxfordjournals.org/ D ow nloaded from 3110 J. Aleksic´ et al. Figure 5. Integrated σLiMa significance as a function of observation time in the 63–158 GeV for the MAGIC stereo system in the case of a GRB event similar to the one reported in this paper. Significance curve evolutions are showed for different starting of observation times after GRB onset: 180 s (green), 600 s (blue), 1100 s (red). In this latter case (inner plot), the foreseen performance for CTA assuming the preliminary sensitivity achieved with MC simulations in the same energy range is showed. The coloured areas shows the assumed 50 per cent systematic errors in the effective area evaluation as explained in Lombardi, Carosi & Antonelli (2013). sensitivity of the future CTA,6 a detection will instead be possible, within the assumed model, even on later time (>1000 s) and higher redshift events. AC K N OW L E D G E M E N T S We acknowledge an anonymous referee for useful comments. We would like to thank the Instituto de Astrofı´sica de Canarias for the excellent working conditions at the Observatorio del Roque de los Muchachos in La Palma. The support of the German BMBF and MPG, the Italian INFN, the Swiss National Fund SNF and the Spanish MICINN is gratefully acknowledged. This work was also supported by the CPAN CSD2007-00042 and MultiDark CSD2009- 00064 projects of the Spanish Consolider-Ingenio 2010 programme, by grant DO02-353 of the Bulgarian NSF, by grant 127740 of the Academy of Finland, by the DFG Cluster of Excellence ‘Origin and Structure of the Universe’, by the DFG Collaborative Research Centers SFB823/C4 and SFB876/C3 and by the Polish MNiSzW grant 745/N-HESS-MAGIC/2010/0. The Fermi-LAT Collaboration acknowledges support from a number of agencies and institutes for both development and the op- eration of the LAT as well as scientific data analysis. These include NASA and DOE in the United States; CEA/Irfu and IN2P3/CNRS in France; ASI and INFN in Italy; MEXT, KEK and JAXA in Japan; and the K. A. Wallenberg Foundation, the Swedish Research Coun- cil and the National Space Board in Sweden. Additional support from INAF in Italy and CNES in France for science analysis during the operations phase is also gratefully acknowledged. This research is partially supported by NASA through the Fermi Guest Investi- gator Grants NNX09AT92G and NNX10AP22G. This work made use of data supplied by the UK Swift Science Data Centre at the University of Leicester. 6 obtained from http://www.cta-observatory.org/ctawpcwiki/index.phpWP_ MC#Interface_to_WP_PHYS. R E F E R E N C E S Abdo A. A. et al., 2009c, ApJ, 706, L138 Acciari V. A. (VERITAS Collaboration) et al., 2011, ApJ, 743, 62 Achterberg A., Gallant Y. A., Kirk J. G., Gauthmann A. W., 2001, MNRAS, 328, 393 Afonso P., Kru¨hler T., Klose S., Greiner J., 2009, GCN Circ., 8771 Aharonian F. et al., 2009, ApJ, 690, 1068 Albert J. et al., 2006, ApJ, 641, 9 Albert J. et al., 2007, ApJ, 667, 358 Albert J. et al., 2008, ApJ, 674, 1037 Aleksicˇ J. et al., 2010, A&A, 517, 5 Aleksicˇ J. et al., 2012, Astropart. Phys., 35, 435 Aliu E. et al., 2008, Science, 322, 1221 Aliu E. et al., 2009a, Astropart. Phys., 30, 293 Altwood W. B. et al., 2009b, ApJ, 697, 1071 Amati L., Frontera F., Guidorzi C., 2009, A&A, 508, 173 Bo¨ttcher M., Dermer C. D., 1998, ApJ, 499, 131 Band D. et al., 1993, ApJ, 413, 281 Band D. L. et al., 2009, ApJ, 701, 1673 Breiman L., 2001, Mach. Learn., 45, 5 Cenko S. B., Rau A., Salvato M., 2009, GCN Circ., 8773 Chandra P., Frail D. A., 2009, GCN Circ., 8779 Costa E. et al., 1997, Nature, 387, 783 Covino S., 2007, Science, 315, 1798 Covino S., Garczarczyk M., Galante N., Gaug M., Antonelli A., Bastieri D., Campana S., Longo F., Scapin V., 2009a, in Bastieri D., Rando R., eds, AIP Conf. Proc. Vol. 1112, Science with the New Generation of High Energy Gamma-Ray Experiments. Am. Inst. Phys., New York, p. 172 Covino S., D’Avanzo P., Antonelli L. A., 2009b, GCN Circ., 8763 Covino S. et al., 2010, A&A, 521A, 53 Dingus B. L., 1995, A&SS, 231, 187 De Ugarte Postigo A., Jakobsson P., Malesani D., Fynbo J. P. U., Simpson E., Barros S., 2009a, GCN Circ., 8766 De Ugarte Postigo A., Blanco L., Castro-Tirando A. J., 2009b, GCN Circ., 8772 Domı´nguez A. et al., 2011a, MNRAS, 410, 2556 Domı´nguez A. et al., 2011b, in Rieger F. M., van Eldik C., Hofmann W., eds, Proc. 25th Texas Symposium on Relativistic Astrophysics, Extragalactic Background Light Inferred from AEGIS Galaxy SED-type Fractions. p. 228 ( PoS(Texas 2010)228) Evans P. A. et al., 2010, A&A, 519, A102 Fan Y. Z., Piran T., 2008, Frontiers Phys. China, 3, 306 Fomin V. P. et al., 1994, Astropart. Phys., 2, 151 Fragile P. C., Mathews G. J. et al., 2004, Astropart. Phys., 20, 591 Franceschini A., Rodighiero G., Vaccari M., 2008, A&A, 487, 837 Fynbo J. P. U. et al., 2009, ApJS, 185, 526 Garczarczyk M. et al., 2009, preprint (arXiv:0907.1001) Gaug M. et al., 2009a, GCN Circ., 8816 Gaug M. et al., 2009b, Proc. 31st ICRC, MAGIC Observation of GRB090102, preprint (astro-ph/0907.0996) Gehrels N. et al., 2004, ApJ, 611, 1005 Gehrels N., Ramirez-Ruiz E., Fox D. B., 2009, ARA&A, 47, 567 Gendre B. et al., 2010, MNRAS, 405, 2372 Ghisellini G., 2010, in Comastri A., Angelini L., Cappi M., eds, AIP Conf. Proc. Vol. 1248, X-Ray Astronomy 2009: Present Status, Multi- Wavelength Approach and Future Perspectives. Am. Inst. Phys., New York, p. 45 Golenetskii S. et al., 2009, GCN Circ., 8776 Gomboc A. et al., 2008, ApJ, 687, 443 Gould R. J., Schrder G., 1966, Phys. Rev. Lett., 16, 252 Granot J. et al., 2010, Proc. Ital. Phys. Soc. Vol. 102, The Shocking Universe – Gamma-Ray Bursts and High Energy Shock phenomena, Ital. Phys. Soc., p. 177 Gupta N., Zhang B., 2007, MNRAS, 380, 78 Hillas A. M., 1985, in Jones F. C., ed., Proc. 19th ICRC, Vol. 3, p. 445 Hurley K. et al., 1994, Nature, 372, 652 Inoue S. et al., 2013, Astropart. Phys., 43, 252 at U niversity of Turku on February 2, 2015 http://m nras.oxfordjournals.org/ D ow nloaded from MAGIC observation of GRB 090102 afterglow 3111 Klebesadel R. W., Strong I. B., Olson R. A., 1973, ApJ, 182, L85 Klotz A., Gendre B., Boer M., Atteia J. L., 2009, GCN Circ., 8761 Kneiske T. M., Bretz T., Manheim K., Hartmann D., 2004, A&A, 413, 807 Levan A. J. et al., 2009, GCN Circ., 8856 Lombardi S., Carosi A., Antonelli L. A., 2013, Proc. 33rd ICRC, preprint (arXiv:1307.3450) Malesani D., Fynbo J. P. U., de Ugarte Postigo A., Gorosabel J., Barros S., Simpson E., 2009, GCN Circ., 8780 Mangano V. et al., 2009a, GCN Rep., 192, 1 Mangano V. et al., 2009b, GCN Circ., 8762 Melandri A. et al., 2010, ApJ, 723, 1331 Me´sz´aros P., 2006, Rep. Prog. Phys. 69, 2259 Mundell C. G. et al., 2007, Science, 315, 1822 Nikishov A. I., 1962, Sov. Phys. JETP, 14, 393 Panaitescu A., Kumar P., 2000, ApJ, 543, 66 Panaitescu A., Kumar P., 2002, ApJ, 571, 779 Perley D. A. et al., 2008, ApJ, 672, 449 Pe’er A., Waxman E., 2005, ApJ, 633, 1018 Rolke W. A., Lo´pez A. M., Conrad J., 2005, Nucl. Instrum. Meth., A551, 493 Sakamoto T. et al., 2009, GCN Circ., 8769 Sari R., Esin A. A., 2001, ApJ, 548, 787 Schweizer T. et al., 2010, 38th COSPAR Scientific Assembly (E110-0016- 10), Status of the MAGIC Telescopes Stecker F. W., Scully S. T., 2008, A&A, 478, L1 Stecker F. W., Malkan M. A., Scully S. T., 2006, ApJ, 648, 774 Steele I. A., Mundell C. G., Smith R. J., Kobayashi S., Guidorzi C., 2009, Nature, 462, 767 Tam P. H. et al., 2006, Nuovo Cimento B, 121, 1595 Van der Horst A. J., Wijers R. A. M. J., Kamble A. P., 2009, GCN Circ., 8792 Vietri M., 2003, ApJ, 591, 954 van Paradijs J. et al., 1997, Nature, 386, 686 Wei D. M., Fan Y. Z., 2007, Chin. J. Astron. Astrophys., 7, 509 Wijers R. A. M. J., Galama T. J., 1999, ApJ, 523, 177 Yost S. A., Harrison F. A., Sari R., Frail D. A., 2003, ApJ, 597, 459 Zhang B., Me´sz´aros P., 2001, ApJ, 559, 110 This paper has been typeset from a TEX/LATEX file prepared by the author. at U niversity of Turku on February 2, 2015 http://m nras.oxfordjournals.org/ D ow nloaded from